An improved African vultures algorithm for multi-threshold optimization in chest X-ray image segmentation.

Area of Science:

• Medical Imaging
• Computer Vision
• Artificial Intelligence

Background:

• Accurate image segmentation is crucial for medical diagnostics and applications.
• Traditional bi-level thresholding methods (Kapur's, Otsu's) are computationally intensive for multi-level segmentation.
• Existing methods struggle with the complexity of multi-thresholding in medical images.

Purpose of the Study:

• To develop a novel multi-level threshold segmentation method for chest X-ray images.
• To enhance the efficiency and accuracy of medical image segmentation using optimization algorithms.
• To address the computational limitations of traditional thresholding techniques in multi-level scenarios.

Main Methods:

• Proposed an Improved Multi-Objective African Vultures Optimization Algorithm (IMMOAVOA).
• Integrated average partial opposite learning and an in-depth exploration mechanism into the AVOA.
• Developed a multi-objective thresholding model combining Otsu's method and 2D Kapur's entropy.

Main Results:

• IMMOAVOA demonstrated superior performance compared to the original AVOA and other benchmark algorithms.
• The proposed method achieved higher efficiency in multi-threshold image segmentation across various metrics.
• Evaluations on ZDT, DTLZ test functions, and chest X-ray images validated the algorithm's effectiveness.

Conclusions:

• The IMMOAVOA offers a significant advancement in multi-level threshold segmentation for medical imaging.
• This approach improves diagnostic precision through more accurate chest X-ray segmentation.
• The study highlights the potential of advanced optimization algorithms in overcoming limitations of traditional image processing techniques.